Hello! Apache on my new server (had it for about 2 weeks, with no modifications done to it yet) keeps dying every few days or so. Apache error logs don't say much, but it keeps dying at the same hour of day - 4am, server time, ever since it was first started. I asked support about it, but they couldn't find anything in the error logs either, so they insist on doing a hardware test, which would consequently result in even more down time. I'm a bit confused, because I don't see how this could be a hardware problem, especially since it keeps happening like a clockwork, always at the same time...

Furthermore, I found out that Apache access logs were first generated (started logging) on May 20th at 4AM (see the pattern?). I told that to support, but they don't seem to find it relevant enough (go figure) and still insist on proceeding with a hardware test, without commenting on that.

My guess is, something is happening on the server every 24 hours since Apache logs were initially started, which sometimes results in Apache dying... could it be log rotation or something like that that crashes it?

If it's a hardware issue, I'll give testing a go, but I really don't want to endure any additional down time unless it's a must...

Thanks!

Very likely it is a cronjob running at 4am and probably the log rotation.. have a look in your /etc/cron.daily/ directory and try disabling the logrotate script and see if apache crashes next time..

Check your /var/log/httpd/error_log and maybe also /var/log/messages for clues. Also check you have enough disk space for the logs.

Thanks Server Mojo. What bugs me is it doesn't happen every day, but rather every few days... I'll definitely check out the cronjobs.

If you look at your /etc/logrotate.conf (which the logrotate cron job uses) you will see how often the logfiles are actually rotated, probably weekly by default.